Design and Performance Metrics of Enhanced ELBP-HOG based hand gesture detection using SVM classifier

Abstract

In human-computer interfacing, Hand Gesture recognition becomes a tough assignment. The gesture recognition techniques were most commonly used in the human-computer communication system. For real-time applications, it is somewhat difficult to satisfy the requirements under various illumination and background conditions. Here in this proposed system, we have improving hand gesture recognition techniques for various computer vision applications and devices. With the help of various feature descriptors, we use Elliptical Local Binary Pattern with Histogram of Oriented Gradients individually, which can concentrate texture and localized portion feature extraction respectively so that the performance can be improved compared with the other systems. Here in the proposed system, the combination of ELBP and HOG can accurately detect various hand poses and we achieved a detection rate that is approximately exceptional compared with other systems. Finally, we used the Support Vector Machine as a classifier for better and accurate results.

Keywords- Gesture, Robustness, Human Computer Interaction, Elliptical Local Binary Pattern-ELBP, HOG-Histogram of Oriented Gradients, SVM-Support Vector Machine